Andreessen Horowitz

A structured curriculum for engineers who have never thought about sales, covering channels, segmentation, quota, comp and coverage in bite-sized pieces. Good for building vocabulary fast so GTM conversations stop being opaque.

All Things Sales! 16 Mini-Lessons for Startup Founders

On Andreessen Horowitz by Peter Levine 16 short videos

Casado's argument that the modern B2B winners combine bottom-up adoption with a sales layer is the clearest framing of why 'which motion are we' is usually the wrong question. Uses Dropbox, Twilio and Atlassian to show the shape.

Growth, Sales, and a New Era of B2B

From Andreessen Horowitz by Martin Casado ~15 min read

• Casado argues the biggest shift in B2B is go-to-market, not technology: bottom-up growth stapled to a sales motion.
• SurveyMonkey turned 60M registrations into a 1.2 billion dollar business at 3 dollars a month.
• Roughly 70 percent of open source companies hit the give-away-versus-sell problem, and most fail it.
• Adding a sales engine can erode the organic growth that made the product spread in the first place.

Maps tiers to how adoption actually spreads (casual user, professional, team, organisation) rather than to a feature matrix, using Notion, Databricks, GitHub and dbt. The best argument for delaying monetisation until usage patterns are legible.

Bottom Up Pricing & Packaging: Let the User Journey Be Your Guide

From Andreessen Horowitz by Jennifer Li and Martin Casado long read

• Across 50 companies studied, the user journey runs casual/freemium, then individual pro, then team, then organization.
• Grow the user base before monetizing: Databricks charged nothing for three years.
• GitHub's 5,000 dollars for 20 seats worked as a rate limiter on inbound demand, not a revenue maximizer.
• Enterprise overtook self-serve around 10M ARR at one infra company and was 70 percent of revenue by 15M.

Five questions you can answer this afternoon from data you already have, including the underused one about what objections churned customers gave you. Written by a16z's GTM network partner, so it is built for operators rather than theorists.

A Framework for Defining and Refining Your ICP

From Andreessen Horowitz by Michael King ~10 min read

• A usable ICP answers nine dimensions including size, geography, industry, job titles, technologies in use, and buyer behaviours.
• Triangulate best customers by fastest revenue growth, shortest sales cycle, and best retention, not revenue alone.
• Mine churned and lost deals for recurring objections: that is where the ICP boundary actually sits.
• ICP is not set and forget, it moves as the product and market do.

The bounding method for a first price (cost to serve as the floor, cost of a heavy power user as the ceiling) plus the advice to keep pricing changeable in the product. Written for founders who are about to underprice, which is most of them.

The Price Is Right: And for Early-Stage SaaS Companies, It Needs to Be

From Andreessen Horowitz by Steven Sinofsky long read

• Founders underprice by pricing against thousands saved when the product actually saves hundreds of thousands.
• In enterprise direct sales, 50 percent or more off list is not unusual and is often required.
• The SaaS norm is a 3 by N matrix: three plans across N attributes, which can imply dozens of real price points.
• Wait until the last possible moment to publish price, because post-launch price changes are never read as good news.

The clearest statement of why per-seat pricing breaks when software does the work, with the support agent example (115 dollars per seat per month becomes a price per resolved ticket) and the variable cost problem underneath AI products.

AI Is Driving A Shift Towards Outcome-Based Pricing

From Andreessen Horowitz by Ivan Makarov, James da Costa and Bobby Pinero short read

• When AI resolves the ticket, Zendesk's 115 dollars per seat per month for support agents stops making sense.
• Three shifts: software becomes labor, the seat stops being the atomic unit, and model API costs make COGS variable.
• Decagon sells both per conversation (usage) and per resolution (outcome) as parallel options.
• AI-native companies price on usage or outcomes while incumbents mostly hold per-seat or bundled models.

Before you can pick a side you need the vocabulary: liquidity, constrained side, multi-tenanting, matching architecture. This is the reference that makes the rest of the marketplace literature readable.

The Marketplace Glossary

From Andreessen Horowitz by Li Jin 20 min read

• Marketplace take rates run from low single digits to the mid-30s, driven by fragmentation, substitutes, and operational value added.
• Liquidity is measured concretely: fill rate (match rate), market depth, and time to find a match.
• Double commit marketplaces need both sides to opt in for a match, which is why Craigslist-style back and forth leaks transactions.

Once you accept that supply is usually the side to define first, this forces the next decision: are you going for everyone, for exclusives, or for a curated set. Each answer produces a completely different supply ICP.

Marketplace Supply Strategy: Comprehensive, Exclusive, or Curated

From Andreessen Horowitz by Casey Winters and Anne Lewandowski 18 min read

• Three supply strategies: comprehensiveness, exclusivity, and curation, and you have to pick one deliberately.
• Being neither comprehensive nor exclusive where variety of supply matters is almost always fatal (Oyster, Tidal).
• Comprehensive does not mean 100 percent coverage, it means enough that users come to you instead of Google.
• Curation works for a niche (HotelTonight optimising for speed and price) but is hard to defend against a well-funded rival.

The best source on stopping the two motions fighting, with real penetration triggers and named cases where it went wrong, including GitHub losing a deal and Slack separating product teams.

The "$20M to $500M" Question: Adding Top Down Sales

From Andreessen Horowitz by Sarah Wang and David George About 15 min read

• The signal to layer in top-down sales is 20 to 30 million dollars of ARR from the bottom-up motion.
• Twilio watched for two or three departments using the product; Dropbox's tipping point was 3 to 10 percent employee penetration.
• Two preconditions before hiring: the bottom-up flywheel works, and the business (not the user) is asking for more.
• Individuals do not buy enterprise features, businesses do, so the product roadmap has to change too.